dgamboa / lc_assignment

MongoDB to SQL Translator

The purpose of the translator is to take in a MongoDB query as input, and return the same query in SQL as the output.

Running the Tests

From the root directory in the terminal, run python -m unittest tests/tests.py.

Running the Translator

Create a sample file in the project's root directory. From the sample file, import main with from translator import main, and call the main.translate function passing in a string containing a MongoDB query: main.translate(query: str) -> str. The function will return a string with the MongoDB query translated to SQL.

Alternatively, import main in the REPL with from translator import main and run the function directly from the terminal.

Engineering Specs

• The translator supports the following MongoDB methods:

db.find()

• The translator returns a simple error message if passed any MongoDB methods not supported.
• The translator supports the following MongoDB operators:

$or
$and
$lt
$lte
$gt
$gte
$ne
$in

Design Decisions

• String as the input / output data structure
  • Queries consist of unicode characters best represented as strings.
• File structure and testing
  • After researching a few projects, I followed the most common structuring guidelines.
  • The file was original an executable script than I refactored into modules to incorporate testing.
• Data structures and algorithms
  • General approach:
    • My primary focus was to build a working translator.
    • The intent was to get a functional MVP that could be later refactored for algorithmic efficiency.
    • I chose to prioritize the delivery guideline (3-4 hours) and simplicity over more sophisticated algorithms that might take more time than intended.
  • Specific choices:
    • The most time-intensive operations are performed by the helper functions that search for characters by looping through strings, or build strings based on a subset of characters specified by indexed locations.
    • The majority of these utilize native functions that search for characters in O(n) time (e.g. find, split, and replace).
    • There are a few instances of custom character-by-character search where necessary (mainly, for instances where arguments need to be split but depend on the sequence of braces or brackets).
    • Some of these may be optimized by incorporating more performant search algorithms.
• Opportunities for future improvement:
  • The logic in several of the helper functions overlaps. This logic should be abstracted into a separate function and leveraged to achieve a DRYer module (e.g. The whereList and whereAndFieldsSeparator functions both use brace counting functionality).
  • It would be useful to implement unit tests for all the helper functions before refactoring them to automate testing for bugs that might be introduced from the changes.
  • My experience building Python packages is limited, so I would want to check with a more experienced developer to incorporate best practices and make imports more robust.

I've included time allocation information in the docs/tracker.md file for additional resolution into my work process